I don't have access to the full report (roll on open access!), but I'm confused by this:

"Of the 3,895 men and 4,148 women who reported drinking alcohol as adults, those with higher average scores on childhood mental ability tests were also more likely to have indications of alcohol problems in adulthood"

How does the picture change if you include the teetotalers? Did they plan from the start to exclude the non-drinker, or is that just being repeated because it's a more interesting figures?